Nikon LS-9000 vs Imacon 848: Resolution & Shadow Detail

[mauro_franic]

Rishi, why fluid mounted and not the Nikon glass holder?

[mauro_franic]

Also, don't forget, when you go to medium format, the Coolscan 9000 will outresolve the Imacon.

[rishij]

Yeah that IS weird. Does anyone know why?

 

I mean, I understand why if the lens is projecting the entire width of a 35mm frame onto the 8000 pixel CCD, then also projecting, through a lens (doing the magnification), a larger format onto the same CCD will result in lower resolution. Sure.

 

But why does this not happen on the LS-9000? Is it as simple as: because when scanning 35mm film, the LS-9000 only uses a portion of its CCD, whereas with the largest format, it uses the full length of the CCD, thereby maintaining resolution?

 

For that matter, DOES the Imacon vary the magnification by the lens to project the 35mm frame fully onto the 8000pixel CCD, then lower the magnification to project a medium format frame onto the same CCD, etc.?

 

Also, let's not forget that multi-sampling significantly reduces noise in dense areas of slide scans. The Imacon shows lots of little green pixels amongst shadows due to its lack of multi-sampling. So does the Nikon at 1x, but not at 4x or 8x.

 

Rishi

[mauro_franic]

"For that matter, DOES the Imacon vary the magnification by the lens to project the 35mm frame fully onto the 8000pixel CCD, then lower the magnification to project a medium format frame onto the same CCD, etc.? "

 

Yes.

 

 

"But why does this not happen on the LS-9000? "

the CCD on the Coolscan has 10,000 pixels and the Imacon has 8,000 pixels. On 120 film the Coolscan captures 8,964 pixels across.

[rishij]

Cool. Thanks Mauro. Can't believe I didn't know that.

 

Since you're a LS-9000 pro, do you find the 1:1 scans on the LS-9000 I made up to snuff?

 

I do believe there is some softness due to the fact that the shot was taken from a bridge, as Vijay pointed out.

 

I'll be back in the next couple of hours to post upsized 8000ppi scan comparisons.

 

Rishi

[mauro_franic]

<p>I think your scans (latest ones) look very good.  There is probably some softness in the film itself but not intrusive.   The difusser on the 9000 does introduce some softness but it retains all the detail - this way you can sharpen and get the best of both worlds.<br>

I don't think there is a better alternative for medium format short of a drum scan.<br>

 </p>

[david_yang2]

<p>Hi there,<br>

I am a total newbie to scanning. I owned a Coolscan V for a long time but never went beyond tweaking with some scanning of old 35mm film before focusing most of my time on DSLRs. Recently I have been moving back to film to get better color and exposure latitude. However rather than going back to 35mm I am shooting 6X6, 6X9 and 6X17. I recently sold my Coolscan V and got myself a Coolscan 9000 so that I can scan the 6X6 and 6X9. I do not have the the glass holder and am interested in knowing what is wet mounting as discussed here? I have visions of dipping film into water . . . also, any suggestion as to scanning 6X17 on a Coolscan 9000? Thanks!<br>

David</p>

[stwrtertbsratbs5]

<p>I use Lumina fluid from Scan Science. Works great! A lot of info is here:</p>

<p>http://scanscience.com/</p>

<p>The proprietor, Julio, is a retired chemist and he's more than happy to talk with potential clients. Give him a call if you have questions.</p>

[matthew_deegan1]

<p>Hi Rishi (and everyone)<br>

appreciate all the work you have done .What a thread!<br>

Rishi, i just wonder a little at your comparison approach. the nikon 9000 has to suffer when the imacon can scan at twice its resolution and then be downsampled to 4000 dpi thereby reducing noise. likewise to upsample the 9000 to 8000dpi seems a bit of a stretch and you should be introducing noise with that exercise too. the two machines have different characteristics and i am wondering about the apples and oranges syndrome. as well, you did say that you had to focus the nikon 9000 a few times and you were not bowled over with that tedium.I did some<strong> ColorLab</strong> testing on the Imacon 343 in terms of color management calibration(Hutch) and surprisingly got better evs results with Imacons own color management system as compared to color produced by those beautiful Hutch targets.I contacted Hasselblad and they were of the view that their own color management system was just right for the Imacon and hard to beat.. I was surprised and a bit disappointed as i am itching to actually use Hutch targets. With that in mind i am trying to combine Silverfast and Hutch( medium format target 6x7 cm) with the 9000 glass set up.. I do not know how to do it...Silverfast are not interested. Can you guys help me please? I have Silverfasts own target setup but i have a hunch that Hutch targets are the business.<br>

Matt</p>

[rishij]

<p>Hi Matt,</p>

<p>My conclusions, in the end, were that the Nikon LS-9000 held up quite well to the Imacon 848. In some cases, it performed better (sharper edge-to-edge scans). This thread is confusing b/c my initial tests were completely off due to a focusing problem with Vuescan.</p>

<p>Now, what software were you using to make the profiles off of the Hutch target scan? I make profiles based off of IT8 and/or Hutch scans... my issue with the Hutch 35mm target is that it has so much pepper grain that, in combination with the smaller sizes of color patches, I worry that the color patches are 'contaminated' by the effects of the black pepper grain. Anyway, I haven't done any deltaE tests on my profiles, but in general I'm quite pleased with the profiles generated with basICColor (trial) off of IT8 scans. I plan on moving back to using LPROF once I can get it to work again under OS X 10.6. </p>

<p>As for my previous claim that Focal Point's Anti-Newton glass (1mm) works well with the Nikon LS-9000... I have to retract my statement. I recently tried some scans with the AN glass, and after IR cleaning, the pattern does show up faintly every now and then in portions of the image. This could be due to IR light fall off near the edges (the pattern mostly shows up on outer portions of the image), or perhaps the LS-9000 light source is not quite diffuse enough (as, say, a light source with a Scanhancer in the optical path). Further testing necessary.</p>

<p>Finally, I'm not quite sure what it is you're trying to do/having trouble with. Please elaborate.</p>

<p>Rishi</p>

[matthew_deegan1]

<p>Hi Rishi<br>

I like Silverfast Ai and want to calibrate a Nikon 9000 with a Hutch 6x 7 cm scanner target.. problem is Silverfast profiler will not see Hutch target's patches as it is it8 compliant and Hutch is not...I have xrite but not for a 6x7 target...I bought BasicColor ICC Profiling software a few years ago which as you know well will take a 6x7 Hutch Target ...BasciColor( I can only surmise)somehow messed up the Hutch TDF data file and made life difficult.They were not helpful and Mr Hutcheson and BasicColor left me to my own devices. I swopped the header info in the TDF from another data file and got the Hutch target working !. That experience was to profile a Imacon 343 but Delta E testing sadly demonstrated that Imacons own color management system for the 343 was superior to Hutch 6x7 profiling...btw i also found that Gamma settings tweaks on the scanner improved Delta E..I ended up using 2.9 gamma with a Silverfast it8 profile...on the Nikon 9000 ..nikon told me that their approach was to vary scanner gamma depending on film densities, which is really cool....anyway starting out,i learnt that best scanner gamma was to keep to your working space gamma( 2.2 gamma for adobe 1998) but delta testing for me showed scanners have a mind of their own. Rishi, Ill dig out the Hutch 6xz7 target and tdf for profiling the Nikon 9000(from BasicColor) . I would appreciate knowing what software company you found that had a simple approach to Hutch medium format targets..i do not feel happy with the experience I have had so far in anything to do with Hutch profiling but i will keep trying. And they are expensive to say the least<br>

Thanks for being so forthcoming with help and information<br>

Matt</p>